The present invention generally relates to an image processing unit and more particularly to an image processing unit in which gray scale image information is converted into binary information by half-tone processing.
In various image processing units, when gray scale image information obtained by reading the intensities in an image such as a photograph is converted into binary information, the binary information is generally obtained by half-tone processing. The Dither method or the error diffusion method are known, for example, as such half-tone processing method.
When a mesh picture, which is, for example, seen in a printed photograph, is converted into binary information by using half-tone processing, moire may be generated. In order to prevent the generation of moire, there is a method in which gray scale image intensity information is smoothed prior to conversion into binary information. In this method, for example, image information comprising 9 pixels (vertical and horizontal 3.times.3) is repeatedly extracted, and image intensity information is smoothed by correcting the pixel positioned in the center of the 9 pixels to the average intensity of the extracted 9 pixels. Conventionally, the average intensity is computed by using a constant number of pixels, such as 3.times.3 pixels, and a uniform smoothing processing is performed over an entire page.
However, because the computation of the average intensity is performed by using an area defined by a constant number of pixels, fine parts of the picture are blurred when the image of a document is scanned with a coarse line density compared to the line density of the mesh picture. On the other hand, when the picture is scanned with a fine line density, the intensity changes are not sufficiently averaged and moire remains. Further, since the smoothing is performed over the entire page, the contours in the picture are unclear.